雷达网信息反馈融合超低空高速强机动目标联合检测与跟踪研究

In recent years, the development of foreign air penetration weapons presents the trend of extreme low-altitude, high speed as well as strong maneuver, and brings a major challenge to existing radar monitoring technology. Conventional radar monitoring methods are more suitable for those high-altitude target detecting and tracking problems with long radar detection range, high detection rate and clear radar echo. However, for extreme low-altitude detecting and tracking problems with short radar detection rang, low detection rate and fuzzy radar echo, few effective methods are available. This project will take extreme low-altitude high speed and strong maneuvering targets as research objects, and will research on the ground-based radar network cooperative detecting and tracking methods. The main research contents are as follows: ① Research on various kinds of extreme low-altitude flying modes, and confirm the key kinematic parameters and establish relative model library; ② Analyze the fuzziness and sparseness of radar echo from the perspective of statistics, and simulate the radar observation data with Monte Carlo methods; ③ Realize the radar network plot fusion cooperative detection and localization algorithm by adopting the random finite set technique, and realize the radar network multiple-model strong maneuvering target tracking algorithm and the observation error adaptive estimating algorithm by adopting the information feedback fusion technique; ④ Demonstrate the effectiveness of the proposed algorithms in the simulation works. By realizing the techniques of radar network plot fusion and multiple-model information feedback fusion, the results can improve the time sensitivity, robustness and accuracy of extreme low-altitude monitoring and provide benchmarks for further research in relative detection and tracking applications.

近年来国外空中突防武器研制呈现超低空、高速和强机动的发展趋势，给现有雷达监测技术带来重大挑战。常规雷达监测方法较适于雷达视距较长、检测率较高和回波数据较清晰的中、高空目标检测与跟踪；而对雷达视距短、检测率低和回波数据模糊的超低空检测与跟踪问题则效果不佳。本项目以超低空高速强机动目标为对象，开展地基雷达网联合检测与跟踪方法研究。其内容包括：①研究各种超低空运动模式、确定关键运动参数、建立相关模型库；②从统计角度分析雷达回波模糊性和稀疏性、用蒙特卡洛方法模拟雷达观测数据；③采用有限集统计技术实现雷达网后验概率假设密度融合联合检测与跟踪算法，采用信息反馈融合技术实现雷达网多模型强机动目标跟踪算法和雷达观测误差自适应估计跟踪算法；④仿真验证算法有效性。以此实现雷达网后验概率假设密度融合和雷达网多模型信息反馈融合技术，提高超低空监测的时敏性、鲁棒性和精度，为进一步研究相关检测与跟踪技术提供参考。

近年来国外空中突防武器研制呈现超低空、高速和强机动的发展趋势，给现有雷达监测技术带来重大挑战。常规雷达监测方法较适于雷达视距较长、检测率较高和回波数据较清晰的中、高空目标检测与跟踪；而对雷达视距短、检测率低和回波数据模糊的超低空检测与跟踪问题则效果不佳。本项目以超低空高速强机动目标为对象，开展地基雷达网联合检测与跟踪方法研究。主要完成了以下研究内容：①建立了低空高速强机动目标的运动模型库；②对地基雷达网的观测数据进行了模拟仿真；③针对雷达观测误差分布不确定问题，提出了观测误差自适应估计目标跟踪算法；④针对强机动目标跟踪问题，提出了误差歧义分解变结构多模型机动目标跟踪算法；⑤针对多雷达GM-PHD信息的融合问题，提出多传感器匹配融合GM-PHD目标跟踪算法；⑥针对多雷达融合跟踪信息对融合顺序敏感的问题，提出多传感器自适应量测迭代更新GM-PHD目标跟踪算法；⑦针对多雷达多机动目标联合检测与跟踪问题，提出多传感器多模型融合GM-PHD目标跟踪算法；⑧搭建了雷达网超低空高速强机动目标联合检测与跟踪仿真验证平台来验证和优化所提算法。通过项目研究完成了预定的研究内容。解决了申请书所提的两个科学问题：①通过后验高斯核的匹配融合和多传感器GM-PHD信息的迭代滤波这两种方式，解决了雷达网后验PHD离散化近似融合计算的问题。②通过构建基于OSPA的全局一致性度量，在匹配融合和迭代更新两种方式下，实现了多传感器估计信息质量的在线评反馈评估方法，达到了优化多传感器估计信息融合顺序的目的，从而解决了雷达网的贝叶斯反馈融合问题。本项目研究所提算法可用于国防和民用领域中的多传感器多目标联合检测与跟踪系统，帮助提高对弱小目标的检测概率以及对机动目标的跟踪精度。